基于机器视觉的洞库类目标自动识别技术  

Research on Cave Target Automatic Recognition Technology based on Robot Vision

作　　者：张辉[1,2] 罗庆生[2] 赵小川[2] 罗霄[2] 高彩[2] 

机构地区：[1]北京理工大学机械与车辆工程学院,北京100081 [2]北京理工大学宇航科学技术学院,北京100081

出　　处：《火力与指挥控制》2010年第9期140-142,146,共4页Fire Control & Command Control

基　　金：省部级重大基金资助项目(1020020220606)

摘　　要：针对传统的洞库类目标识别方法所存在的不足,设计了一种基于可见光/红外线探测器和嵌入式DSP的洞库类目标主动探测和自动识别装置,根据洞库外观的结构特征建立了洞库识别的数学模型,提出了基于Hough变换的洞库类目标自动识别算法,并将洞库识别数学模型与Hough变换的处理流程有机结合起来,实现了不同外观结构的洞库类目标的正确识别。实验结果表明该洞库类目标识别算法具有较强的鲁棒性,可以有效地抑制噪声的影响,对由于预处理参数设置不合理所造成的洞库类目标边缘成为不连续曲线的现象也有良好的自适应性和自判断性。In order to overcome the shortcomings of traditional cave target recognition methods,a cave target automatic recognition equipment based on visible light/infrared sensor and embedded DSP has been designed.The cave targets are modeled according to their structure characteristics.A cave target automatic recognition algorithm based on Hough transform is proposed.Experimental results demonstrate that the cave target automatic recognition algorithm proposed in this paper has good robustness,which can restrain noises effectively.This method also has good adaptability towards discontinuous target edge caused by unreasonable pretreatment parameters setting.

关 键 词：目标识别 机器视觉 洞库类目标 HOUGH变换 

分 类 号：TP75[自动化与计算机技术—检测技术与自动化装置]